CN109657664A - A kind of recognition methods, device and the electronic equipment of license plate type - Google Patents
A kind of recognition methods, device and the electronic equipment of license plate type Download PDFInfo
- Publication number
- CN109657664A CN109657664A CN201710946300.2A CN201710946300A CN109657664A CN 109657664 A CN109657664 A CN 109657664A CN 201710946300 A CN201710946300 A CN 201710946300A CN 109657664 A CN109657664 A CN 109657664A
- Authority
- CN
- China
- Prior art keywords
- character
- license plate
- target
- feature
- distributed architecture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
- G06V10/23—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on positionally close patterns or neighbourhood relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/158—Segmentation of character regions using character size, text spacings or pitch estimation
Abstract
The embodiment of the present application provides recognition methods, device and the electronic equipment of a kind of license plate type.The described method includes: obtaining license plate image to be identified, the license plate area in the license plate image to be identified is positioned;Character recognition is carried out to the license plate area, obtains character identification result;According to the character identification result, the corresponding target character feature of the license plate area is determined;The corresponding target license plate type of the target character feature, the license plate typelib, for storing the corresponding relationship of character feature Yu license plate type are determined from preset license plate typelib;According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.Using scheme provided by the embodiments of the present application, accuracy rate when identification license plate type can be improved in the identical situation of license plate color information of different license plate types.
Description
Technical field
This application involves field of intelligent transportation technology, more particularly to recognition methods, device and the electricity of a kind of license plate type
Sub- equipment.
Background technique
License plate is vehicle " identity card ", is an important information for being different from other motor vehicles.License plate can be divided into
Multiple types.In same country or area, license plate can be divided into common license plate, coach's board, interim listed, Wu Jingche
The types such as board, military license plate.In the area of the open to traffic of multiple countries, what can be belonged to according to license plate is national by license plate
Divide into different types.Identify that the type of license plate has important application in the message context for obtaining vehicle.
It in the related technology, can be according to license plate to be identified when the license plate type to license plate image to be identified identifies
The color characteristic of license plate area in image, from license plate to be identified determining in the corresponding relationship of preset license plate type and color space
The corresponding license plate type of image.Specifically, license plate area can be oriented from license plate image to be identified, the license plate area is counted
The pixel quantity of middle different colours determines the pixel that the corresponding color space of preset license plate type is fallen in the license plate area
Point quantity accounts for the ratio of the total pixel quantity of the license plate area, and the color space which is greater than preset ratio threshold value is corresponding
License plate type is determined as the license plate type of license plate image to be identified.
In general, when each license plate type corresponds to different color spaces, using the recognition methods energy of above-mentioned license plate type
Enough identify the license plate type of license plate image to be identified.But due to there is a large amount of license plate type, each license plate class in practice
The corresponding license plate color information of type is not different from each other, therefore identical for the license plate color information of different license plate types
The accuracy rate of situation, the license plate type identified using the above method is lower.For example, as shown in Figure 1a, the military license plate 1 in China
It is all the license plate of white gravoply, with black engraved characters with police license plate 2, the license plate of Croatia 3 and Slovakia 4 is all the license plate of white gravoply, with black engraved characters, is adopted
Can not accurately distinguish in aforementioned manners license plate image to be identified be military license plate type or police license plate type, also can not be accurate
Distinguishing license plate image to be identified is Croatia's license plate type or Slovakia's license plate type.
Summary of the invention
Recognition methods, device and the electronic equipment for being designed to provide a kind of license plate type of the embodiment of the present application, with
In the identical situation of license plate color information of different license plate types, accuracy rate when identification license plate type is improved.
In order to achieve the above object, the embodiment of the present application provides a kind of recognition methods of license plate type, the method packet
It includes:
License plate image to be identified is obtained, the license plate area in the license plate image to be identified is positioned;
Character recognition is carried out to the license plate area, obtains character identification result;
According to the character identification result, the corresponding target character feature of the license plate area is determined;
The corresponding target license plate type of the target character feature, the license plate class are determined from preset license plate typelib
Type library, for storing the corresponding relationship of character feature Yu license plate type;
According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.
Optionally, the license plate typelib, for storing each layer character feature and the last layer according to tree-like arrangement
Each layer in the corresponding license plate type of character feature, each target character feature and the license plate typelib corresponds;
The described the step of corresponding target license plate type of the target character feature is determined from preset license plate typelib,
Include:
First layer target character feature is determined as to the target character feature of current layer, it will be first in the license plate typelib
All character features of layer are determined as character feature to be selected;
The determining matching character feature with the target character characteristic matching of current layer from the character feature to be selected;
Judge whether the matching character feature is the last layer character feature;
If it is, the corresponding license plate type of the matching character feature is determined as target license plate type;
If it is not, then next layer of target character feature to be updated to the target character feature of current layer, by the license plate class
Next layer of all character features are updated to character feature to be selected in type library, return special from the character to be selected described in executing
In sign the step of the determining matching character feature with the target character characteristic matching of current layer.
Optionally, the target character feature includes at least one of following characteristics: single bilayer license board information, character are total
Quantity, the quantity of all kinds of ocra font ocrs, character distributed architecture.
Optionally, the target character feature includes the character distributed architecture;
It is described according to the character identification result, the step of determining the license plate area corresponding target character feature, packet
It includes:
According to the character identification result, the space position in the license plate area is determined;
According to each character in the space position and the character identification result, the corresponding word of the license plate area is determined
Accord with distributed architecture.
Optionally, described according to the character identification result, the step of determining the space position in the license plate area, packet
It includes:
Determine center in the character identification result between adjacent character region away from;
Determine the center away from the maximum preset quantity center of middle numerical value away from, as target's center away from;
Calculate center of the center away from in addition to the target's center away from other than away from average value, as mean center
Away from;
Judge the target's center away from the mean center away from whether meeting default size relation;
If it is satisfied, then the target's center is determined as space position away from the position between corresponding adjacent character region
It sets;
Wherein, the default size relation includes at least one of following situations:
The target's center away from the mean center away from ratio be greater than default fractional threshold;
The target's center away from the mean center away from difference be greater than preset difference value threshold value.
Optionally, after determining the corresponding character distributed architecture of the license plate area, the method also includes:
Judge whether the character distributed architecture meets preset characters distributed architecture;
If met, judge whether the character types in the character identification result are preset characters type, it is described pre-
If character types are character types corresponding with the preset characters distributed architecture;
If being not preset characters type, the character in the character identification result is converted to corresponding with the character
The character of preset characters type obtains updated character identification result.
Optionally, described according to each character in the space position and the character identification result, determine the license plate
The step of corresponding character distributed architecture in region, comprising:
According to each character in the space position and the character identification result, original character distributed architecture is determined;
Judge in preset structural library with the presence or absence of distributed architecture identical with the original character distributed architecture, wherein
The structural library is for storing each character distributed architecture;
If it is present the original character distributed architecture is determined as the corresponding character distribution knot of the license plate area
Structure;
If it does not exist, then the determining and highest character of original character distributed architecture similarity from the structural library
Distributed architecture, as reference character distributed architecture;
Determine the difference character field of the original character distributed architecture Yu reference character distributed architecture;
According to the difference character field, character corresponding with the difference character field in the character identification result is corrected,
Obtain revised character identification result;
According to the space position and revised character identification result, the corresponding character distribution of the license plate area is determined
Structure.
Optionally, described according to the target license plate type, determine the corresponding license plate type of the license plate image to be identified
The step of, comprising:
Judge whether the quantity of the target license plate type is greater than one;
If it is not, then the target license plate type is determined as the corresponding license plate type of the license plate image to be identified;
If it is, determining distinctive mark region from the license plate area according to preset distinctive mark position;
Determine the characteristics of image in the distinctive mark region, and be used to determine license plate by the input of described image feature is preset
The classifier of type;Wherein, the classifier, characteristics of image and vehicle obtained when for being completed according to the classifier training
The corresponding relationship of board type determines the inputted corresponding license plate type of characteristics of image;Described image feature is according to the spy
The feature that the pixel value of different mark region determines;
The license plate type that the classifier is sent is obtained, accessed license plate type is determined as the license plate to be identified
The corresponding license plate type of image.
In order to achieve the above object, the embodiment of the present application provides a kind of identification device of license plate type, described device packet
It includes:
Locating module positions the license plate area in the license plate image to be identified for obtaining license plate image to be identified;
Identification module obtains character identification result for carrying out character recognition to the license plate area;
First determining module, for determining the corresponding target character of the license plate area according to the character identification result
Feature;
Matching module, for determining the corresponding target license plate class of the target character feature from preset license plate typelib
Type, the license plate typelib, for storing the corresponding relationship of character feature Yu license plate type;
Second determining module, for determining the corresponding vehicle of the license plate image to be identified according to the target license plate type
Board type.
Optionally, the license plate typelib, for storing each layer character feature and the last layer according to tree-like arrangement
Each layer in the corresponding license plate type of character feature, each target character feature and the license plate typelib corresponds;
The matching module, comprising:
First determines submodule, will for first layer target character feature to be determined as to the target character feature of current layer
All character features of first layer are determined as character feature to be selected in the license plate typelib;
Matched sub-block, for target character characteristic matching determining and current layer from the character feature to be selected
Match character feature;
First judging submodule, for judging whether the matching character feature is the last layer character feature;
Second determines submodule, is used for when the matching character feature is the last layer character feature, by the matching
The corresponding license plate type of character feature is determined as target license plate type;
Submodule is updated, for when the matching character feature is not the last layer character feature, by next layer of target
Character feature is updated to the target character feature of current layer, and next layer in the license plate typelib of all character features are updated
For character feature to be selected, the matched sub-block is triggered.
Optionally, the target character feature includes character distributed architecture;
First determining module, comprising:
Space determines submodule, for determining the space position in the license plate area according to the character identification result;
Structure determination submodule, for determining according to each character in the space position and the character identification result
The corresponding character distributed architecture of the license plate area.
Optionally, the space determines submodule, comprising:
First center is away from determination unit, for determining the center in the character identification result between adjacent character region
Away from;
Second center is away from determination unit, for determining the center away from the maximum preset quantity center of middle numerical value away from work
For target's center away from;
Mean center away from computing unit, for calculate center of the center away from in addition to the target's center away from other than away from
Average value, as mean center away from;
First judging unit, for judge the target's center away from the mean center away from whether meeting default size and close
System;
Space determination unit is used for when the target's center is away from the mean center away from default size relation is met,
The target's center is determined as space position away from the position between corresponding adjacent character region;
Wherein, the default size relation includes at least one of following situations:
The target's center away from the mean center away from ratio be greater than default fractional threshold;
The target's center away from the mean center away from difference be greater than preset difference value threshold value.
Optionally, described device further include:
Update module, for after determining the corresponding character distributed architecture of the license plate area, judging the character point
Whether cloth structure meets preset characters distributed architecture;If met, judge that the character types in the character identification result are
No is preset characters type, and the preset characters type is character types corresponding with the preset characters distributed architecture;If
It is not preset characters type, then the character in the character identification result is converted into preset characters type corresponding with the character
Character, obtain updated character identification result.
Optionally, the structure determination submodule, comprising:
Originally determined unit, for determining just according to each character in the space position and the character identification result
Beginning character distributed architecture;
Second judgment unit, for judging in preset structural library with the presence or absence of identical as the original character distributed architecture
Distributed architecture, wherein the structural library is for storing each character distributed architecture;
First structure determination unit, for identical with the original character distributed architecture when existing in preset structural library
When distributed architecture, the original character distributed architecture is determined as the corresponding character distributed architecture of the license plate area;
Second structure determination unit, for when there is no identical as the original character distributed architecture in preset structural library
Distributed architecture when, then it is determining from the structural library to be tied with the highest character distribution of the original character distributed architecture similarity
Structure, as reference character distributed architecture;Determine the difference character of the original character distributed architecture Yu reference character distributed architecture
Section;According to the difference character field, character corresponding with the difference character field in the character identification result is corrected, is repaired
Character identification result after just;According to the space position and revised character identification result, the license plate area pair is determined
The character distributed architecture answered.
Optionally, second determining module, comprising:
Second judgment submodule, for judging whether the quantity of the target license plate type is greater than one;
Third determines submodule, when the quantity of the target license plate type is not more than one, by the target license plate class
Type is determined as the corresponding license plate type of the license plate image to be identified;
4th determine submodule, for when the quantity of the target license plate type be greater than one when, according to preset special
Mark position determines distinctive mark region from the license plate area;
Input submodule inputs in advance for determining the characteristics of image in the distinctive mark region, and by described image feature
If for determining the classifier of license plate type;Wherein, the classifier is obtained when for completing according to the classifier training
The corresponding relationship of the characteristics of image and license plate type that obtain determines the inputted corresponding license plate type of characteristics of image;Described image
Feature is the feature determined according to the pixel value in the distinctive mark region;
Acquisition submodule, the license plate type sent for obtaining the classifier, accessed license plate type is determined
For the corresponding license plate type of the license plate image to be identified.
The embodiment of the present application provides a kind of electronic equipment, including processor, communication interface, memory and communication bus,
Wherein, processor, communication interface, memory complete mutual communication by communication bus;
Memory, for storing computer program;
Processor when for executing the program stored on memory, realizes license plate type provided by the embodiments of the present application
Recognition methods.This method comprises:
License plate image to be identified is obtained, the license plate area in the license plate image to be identified is positioned;
Character recognition is carried out to the license plate area, obtains character identification result;
According to the character identification result, the corresponding target character feature of the license plate area is determined;
The corresponding target license plate type of the target character feature, the license plate class are determined from preset license plate typelib
Type library, for storing the corresponding relationship of character feature Yu license plate type;
According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.
The embodiment of the present application provides a kind of computer readable storage medium, storage in the computer readable storage medium
There is computer program, the computer program realizes the identification of license plate type provided by the embodiments of the present application when being executed by processor
Method.This method comprises:
License plate image to be identified is obtained, the license plate area in the license plate image to be identified is positioned;
Character recognition is carried out to the license plate area, obtains character identification result;
According to the character identification result, the corresponding target character feature of the license plate area is determined;
The corresponding target license plate type of the target character feature, the license plate class are determined from preset license plate typelib
Type library, for storing the corresponding relationship of character feature Yu license plate type;
According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.
Recognition methods, device and the electronic equipment of license plate type provided by the embodiments of the present application, can position vehicle to be identified
License plate area in board image carries out character recognition to the license plate area, character identification result is obtained, according to the character recognition knot
Fruit determines the corresponding target character feature of license plate area, determines that target character feature is corresponding from preset license plate typelib
Target license plate type, and according to the target license plate type, determine the corresponding license plate type of license plate image to be identified.Wherein, license plate
Typelib is used to store the corresponding relationship of character feature Yu license plate type.
That is, the embodiment of the present application can determine license plate type according to the character feature of license plate.In different license plates
In the identical situation of license plate color information of type, the character feature of license plate is the feature more richer than the colouring information of license plate,
License plate type can be more accurately identified according to the character feature of license plate.Meanwhile to license plate area progress character recognition, and according to
Character identification result determines the character feature of license plate, and determining character feature can be made more acurrate, to further increase this Shen
Please embodiment provide license plate type identification scheme accuracy rate.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described.It should be evident that the accompanying drawings in the following description is only this
Some embodiments of application for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 a is the different types of license plate exemplary diagram of the identical license plate of colouring information;
Fig. 1 b is the license plate exemplary diagram of European Union area every country type;
Fig. 2 is a kind of flow diagram of the recognition methods of license plate type provided by the embodiments of the present application;
Fig. 3 is a kind of flow diagram of step S204 in Fig. 2;
Fig. 4 be according to character zone determine center away from a kind of schematic illustration;
Fig. 5 is a kind of flow diagram of step S205 in Fig. 2;
Fig. 6 is a kind of structural schematic diagram of the identification device of license plate type provided by the embodiments of the present application;
Fig. 7 is a kind of structural schematic diagram of electronic equipment provided by the embodiments of the present application.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete
Whole description.Obviously, described embodiment is only a part of the embodiment of the application, instead of all the embodiments.Base
Embodiment in the application, those of ordinary skill in the art are obtained all without making creative work
Other embodiments shall fall in the protection scope of this application.
The embodiment of the present application provides recognition methods, device and the electronic equipment of a kind of license plate type, can be in different vehicles
In the identical situation of license plate color information of board type, accuracy rate when identification license plate type is improved.Below by specific implementation
Example, is described in detail the application.
Fig. 2 is a kind of flow diagram of the recognition methods of license plate type provided by the embodiments of the present application.This method application
In electronic equipment.This method specifically comprises the following steps S201~step S205:
Step S201: obtaining license plate image to be identified, positions the license plate area in the license plate image to be identified.
Wherein, license plate image to be identified can be understood as the image comprising license plate area of license plate type to be identified.Wait know
Other license plate image can be the vehicle image captured on road, be also possible to image arbitrarily comprising license plate area, the application
Embodiment is not specifically limited in this embodiment.License plate area can be understood as the image-region where the license plate number of vehicle.
As a kind of specific embodiment, license plate can be divided into multiple types, license plate to be identified according to actual needs
License plate in image may be one in each type.For example, in same country or area, it can be in advance by license plate point
For the types such as common license plate, coach's board, interim listed, People's Armed Police's license plate, military license plate;In the open to traffic of multiple countries
Area (such as European Union area), can previously according to license plate belong to country license plate be divided into different country types.To
The license plate of identification license plate image may be any one in above-mentioned license plate type.
For example, Fig. 1 b, which is shown, divides obtained country according to the national difference that license plate belongs to for the license plate of European Union area
Type.7 national license plates are distinguished with 7 frames in figure, the corresponding country type of these license plates is respectively as follows: Croatia K, Si Luo
Cut down a gram S, Czech J, Bulgarian B, Macedonia M, Hungary X, Switzerland R.
It is positioned in license plate image to be identified after obtaining license plate image to be identified as the electronic equipment of executing subject
License plate area when, can be according to there are the features of a large amount of pixel value trip points to be positioned in license plate area, can also basis
The edge of license plate area is positioned with the feature that other area pixel points have jump.The application to specific position fixing process not
It limits.
Step S202: character recognition is carried out to above-mentioned license plate area, obtains character identification result.
Wherein, above-mentioned character identification result may include character and corresponding character types, character zone etc., character zone
The as character corresponding image-region in license plate area.
Specifically, carrying out character recognition to license plate area, when obtaining character identification result, vertical projection method can be first used
And/or connection domain method is split the first license plate area, obtains Character segmentation as a result, then using preset Character recognizer
Each Character segmentation result is identified, character identification result is obtained.
Step S203: according to above-mentioned character identification result, the corresponding target character feature of above-mentioned license plate area is determined.
Wherein, target character feature includes at least one of following characteristics: single bilayer license board information, character total quantity,
The quantity of all kinds of ocra font ocrs, character distributed architecture.It is that single layer license plate is still double that single bilayer license board information, which can be understood as the license plate,
Layer license plate.All kinds of ocra font ocrs may include the character of letter type and the character of numeric type.Character distributed architecture is understood that
For the structure as corresponding to the space among character number and character.Space among character, referring in license plate area can incite somebody to action
Character is divided into the part of each character field, and first license plate area includes 2 spaces to example in Croatia K type as shown in figure 1, this
Character is divided into 3 sections by a little spaces.In addition, character distributed architecture can also include the character types of each section character.For example, figure
The corresponding character distributed architecture of first license plate area is 3-3 in Hungary X type in 1, and intermediate symbol "-" represents space,
And the corresponding character types of 3-3 structure are letter type-numeric type.
It, can be with when determining the corresponding target character feature of above-mentioned license plate area specifically, according to above-mentioned character identification result
It include: to determine that above-mentioned license plate area is corresponding according to the character and corresponding character types for including in above-mentioned character identification result
The quantity of character total quantity and all kinds of ocra font ocrs;According to the character zone for each character for including in above-mentioned character identification result, really
Determine the corresponding single double-deck license board information of above-mentioned license plate area and character distributed architecture.
Step S204: the corresponding target license plate type of above-mentioned target character feature is determined from preset license plate typelib.
Wherein, license plate typelib, for storing the corresponding relationship of character feature Yu license plate type.In the present embodiment,
Corresponding sample license plate image can be obtained previously according to ready-portioned each license plate type, be mentioned from each sample license plate image
Character feature is taken, establishes the corresponding relationship of character feature Yu license plate type, and the corresponding relationship is stored to license plate typelib.It should
Character feature may include at least one of following characteristics: single bilayer license board information, character total quantity, all kinds of ocra font ocrs number
Amount, character distributed architecture.
Specifically, may include following two mode when establishing the corresponding relationship of character feature and license plate type:
The first, establishes the corresponding relationship between each license plate type and each character feature.
It is illustrated by taking the license plate of the every country type of European Union area as an example below.Referring to every country shown in Fig. 1 b
Corresponding sample license plate image, can establish the corresponding relationship between character feature shown in table 1 and license plate type.
Table 1
Character feature | Country type |
L=0, Ft=6~8, Fe=3~6, Fn=1~4, Fc=2-3-2,2-3-1,2-4-1,2-4-2 | K |
L=0, Ft=7, Fe=4, Fn=3, Fc=2-5 | S |
L=0, Ft=7, Fe=1~2, Fn=5~6, Fc=3-4 | J |
L=0, Ft=7~8, Fe=3~4, Fn=4, Fc=1-4-2,2-4-2,1-6 | B |
L=0, Ft=7~8, Fe=4, Fn=3~4, Fc=2-3-2,2-4-2 | M |
L=-1, Ft=6, Fe=3, Fn=3, Fc=3-3 | X |
L=-1, Ft=6~8, Fe=2~3, Fn=4~6, Fc=2-4,2-5,2-5-1,2-6 | R |
Wherein, L is single dual-layered information, and L=0 expression single layer, L=1 expression is double-deck, and L=-1 indicates that single bilayer has, and Ft is
Character total quantity, Fe are alphabetical total quantity, and Fn is numerical sum amount, and Fc is character distributed architecture.
Second, establish the corresponding relationship between each character feature and license plate type.
Still it is illustrated by taking the license plate of the every country type of European Union area as an example.Referring to every country pair shown in Fig. 1 b
The sample license plate image answered, can establish the corresponding relationship between character feature shown in table 2 and license plate type.
Wherein, the meaning of each letter and table 1 are same in table.
Specifically, when determining the corresponding target license plate type of above-mentioned target character feature from preset license plate typelib,
It can correspond to including following two embodiment:
Mode one: each target character feature is matched with each character feature in preset license plate typelib,
The corresponding license plate type of the character feature that each target character feature can be matched, is determined as target license plate type.It is this
Mode corresponds to license plate typelib shown in table 1.
Mode two: by each target character feature according to it is preset put in order one by one with each word in license plate typelib
Symbol feature is matched, and determines the corresponding license plate type group of each target character feature, includes at least one in the license plate type group
A license plate type;The identical license plate type for including in each license plate type group is determined as target license plate type.
Step S205: according to above-mentioned target license plate type, the corresponding license plate type of license plate image to be identified is determined.
Specifically, directly target license plate type is determined as wait know when the quantity of above-mentioned target license plate type is one
The corresponding license plate type of other license plate image;It, can be according to license plate area when the quantity of above-mentioned target license plate type is greater than one
In special pattern further screened, determine the corresponding license plate type of license plate image to be identified.
As shown in the above, the present embodiment can determine license plate type according to the character feature of license plate.Character feature
It may include single double-deck license board information, character total quantity, the quantity of all kinds of ocra font ocrs, character distributed architecture etc..In different license plates
In the identical situation of license plate color information of type, the character feature of license plate is the feature more richer than the colouring information of license plate,
License plate type can be more accurately identified according to the character feature of license plate.Meanwhile to license plate area progress character recognition, and according to
Character identification result determines the character feature of license plate, and determining character feature can be made more acurrate, to further increase this reality
The accuracy rate of the license plate type identification scheme of example offer is provided.
In order to improve matched accuracy and efficiency, in the another embodiment of embodiment shown in Fig. 2, the license plate
Typelib can be used for storing the corresponding license plate class of each layer character feature and the last layer character feature according to tree-like arrangement
Each layer in type, each target character feature and license plate typelib corresponds.
By taking the license plate of the every country type of European Union area as an example, illustrate character feature and country type in present embodiment
Between corresponding relationship.Referring to the corresponding sample license plate image of every country shown in Fig. 1 b, character shown in table 3 can establish
Corresponding relationship between feature and license plate type.
Table 3
Wherein, license plate list dual-layered information is first layer character feature, and character total quantity is second layer character feature, character point
Cloth structure is third layer character feature and the last layer character feature, and character distributed architecture is corresponding with license plate type.
In the present embodiment, step S204 determines that the target character feature is corresponding from preset license plate typelib
The step of target license plate type, can carry out according to flow diagram shown in Fig. 3, specifically include following steps S204a~
S204e:
First layer target character feature: being determined as the target character feature of current layer by step S204a, by above-mentioned license plate class
All character features of first layer are determined as character feature to be selected in type library.
Step S204b: the determining matching character with the target character characteristic matching of current layer from character feature to be selected
Feature.
Specifically, the determining matching character feature with the target character characteristic matching of current layer from character feature to be selected
When, it may include being determined as the identical character feature of target character feature in character feature to be selected with current layer to match word
Accord with feature.
Step S204c: judge whether above-mentioned matching character feature is the last layer character feature, if it is, executing step
Rapid S204d;If not, thening follow the steps S204e.
Step S204d: the corresponding license plate type of above-mentioned matching character feature is determined as target license plate type.
Next layer of target character feature: being updated to the target character feature of current layer by step S204e, by license plate typelib
In next layer of all character features be updated to character feature to be selected, return to step S204b.
As an example, it is known that target character feature includes: single layer license plate, and character total quantity is 6, character distributed architecture
For 2-4.Wherein, with the sequence of the one-to-one target character feature of each layer in license plate typelib are as follows: first layer be single layer vehicle
Board, the second layer are that character total quantity is 6, and third layer is that character distributed architecture is 2-4.License plate typelib is as shown in table 3, then pressing
When being matched target character feature with license plate typelib according to above-mentioned matching process, it can determine that target license plate type is R.
As it can be seen that each layer character feature can be stored in license plate typelib by the present embodiment according to tree-like put in order,
When matching, each target character feature is sequentially matched according to the sequence of each layer, until it is matched to the last layer character feature, it will most
The corresponding license plate type of later layer character feature is determined as target license plate type.Matching efficiency can be improved in this matching process,
Guarantee matched accuracy simultaneously.
In a kind of embodiment of embodiment shown in Fig. 2, when target character feature includes character distributed architecture, step
S203 determines the corresponding target character feature of above-mentioned license plate area, may include following step that is, according to above-mentioned character identification result
Rapid 1 and step 2:
Step 1: according to above-mentioned character identification result, determining the space position in above-mentioned license plate area.
Wherein it is determined that the space position in license plate area can be understood as determining that space is located at n-th of word in license plate area
Between symbol and m-th of character, it is understood that for the coordinate position for determining space in license plate area.
Specifically, the step 1 may include numerous embodiments, as a kind of specific embodiment, step 1 be can wrap
Include following steps 1a~step 1e:
Step 1a: determine center in above-mentioned character identification result between adjacent character region away from.
Specifically, determine center away from when, may include: the vertical of each character zone in determining above-mentioned character identification result
The abscissa of center line determines the distance between adjacent upright center line, obtain each center away from.
As an example, Fig. 4 provide according to character zone determine center away from a kind of schematic illustration.Square is used in figure
Shape box is shown each character zone (6) totally, and vertical dotted line is the vertical centerline (totally 6) of each character zone, respectively
The distance between a vertical centerline is gone out with lateral arrows line drawing, the length of the lateral arrows line be centered on away from (totally 5).
Step 1b: determine above-mentioned center away from the maximum preset quantity center of middle numerical value away from, as target's center away from.Its
In, preset quantity can be configured according to the actual situation, such as when at most there is 2 spaces in actual license plate, can be incited somebody to action
Preset quantity is set as 2.
Step 1c: calculate center of the above-mentioned center away from in addition to target's center away from other than away from average value, in average
The heart away from.
Step 1d: judge above-mentioned target's center away from mean center away from whether meeting default size relation, if it is satisfied, then
Execute step 1e;If conditions are not met, can not then be handled.
Wherein, above-mentioned default size relation may include at least one of following situations:
Target's center away from mean center away from ratio be greater than default fractional threshold;
Target's center away from mean center away from difference be greater than preset difference value threshold value.
For example, preset ratio threshold value can be taken as 1.1,1.2 or other values.Preset difference value threshold value can be according to license plate word
Symbol center away from assembly average obtain, for example, when characters on license plate center away from assembly average be 10px when, preset difference value threshold
Value can be taken as 2px, 3px or other values.
It is understood that mean center is away from the center being able to reflect between the character not comprising space away from when in target
When the heart is away from mean center away from above-mentioned default size relation is met, the target's center can be determined away from corresponding space position.
When target's center away from quantity be greater than two when, judge above-mentioned target's center away from mean center away from whether meeting
When default size relation, can judge respectively each target's center away from mean center away from whether meeting default size relation.
Step 1e: target's center is determined as space position away from the position between corresponding adjacent character region.
It should be pointed out that each target's center is calculated away from being by two adjacent character zones, therefore can
The position between two adjacent character zones is determined as space position.Specifically, can be by two adjacent character areas
Regional location in domain between the right margin of left character zone and the left margin of right character zone is determined as space position.It can also
To be, by the position between two adjacent character zones it is corresponding be information between which character and which character,
It is determined as space position.
When there are multiple target's centers in the preset quantity target's center determined in step 1b away from meeting in step 1d away from
Rule of judgment then illustrates that there are how many a spaces in the license plate area.
Further, the corresponding target's center in each space position can also be recorded away from and determining each mesh for being recorded
Ratio of the mark center away between, as the information that character distributed architecture includes in character feature.
It should be noted that center in above embodiment by character pitch away from can also be replaced.Character pitch can be with
It is interpreted as the right margin of left character zone and the distance between the left margin of right character zone.
Step 2: according to each character in above-mentioned space position and character identification result, determining the corresponding character of license plate area
Distributed architecture.
Specifically, can determine license plate area according to the sequence of each character in above-mentioned space position and character identification result
The corresponding character distributed architecture in domain;It can also be according to the corresponding character of character each in above-mentioned space position and character identification result
The position in region determines the corresponding character distributed architecture of license plate area.
As an example, when determine space position between the 3rd character and the 4th character, and character recognition knot
6 characters are shared in fruit, can determine that character distributed architecture is 3-3.
For another example, when determining that space position is [20,25], and the corresponding character zone of each character in character identification result
Respectively [1,5], [7,12], [14,19], [26,31], [33,38] when [40,45], can determine that character distributed architecture is 3-
3。
Alternatively, step 2 determines i.e. according to each character in above-mentioned space position and character identification result
The step of license plate area corresponding character distributed architecture, it may comprise steps of 2a~step 2g:
Step 2a: according to each character in above-mentioned space position and character identification result, original character distributed architecture is determined.
Step 2b: judging to whether there is distributed architecture identical with the original character distributed architecture in preset structural library,
If it is present executing step 2c;If it does not exist, then executing step 2d.
Wherein, above structure library is for storing each character distributed architecture.As an example, above structure library can be
The structural library of each character distributed architecture in storage table 3.
Step 2c: the original character distributed architecture is determined as the corresponding character distributed architecture of license plate area.
Step 2d: the determining and highest character distributed architecture of original character distributed architecture similarity from the structural library is made
For reference character distributed architecture.
Specifically, when determining the similarity between the character distributed architecture in structural library and original character distributed architecture,
It may include numerous embodiments.It, can be according to the space in original character distributed architecture as a kind of specific embodiment
Quantity filters out the partial character distributed architecture with the space quantity from structural library;Then it is distributed and is tied according to original character
The character number of each character field in structure is filtered out with original character distributed architecture from partial character distributed architecture with identical
The largest number of subdivision character distributed architectures of the character field of character number;Determine original character distributed architecture and subdivision character
The different character field of distributed architecture character number, by the smallest subdivision character distributed architecture of the character number gap of character field,
It is determined as and the highest character distributed architecture of original character distributed architecture similarity.
For example, original character distributed architecture is 3-4-2, then according to each character distributed architecture in table 3 it is found that having
The partial character distributed architecture of two spaces includes: 2-3-1,2-3-2,2-4-1,1-4-2,2-4-2,2-5-1.These partial words
With original character distributed architecture there is the character field number of identical characters number to be respectively as follows: 0,1,1,2,2,0 in symbol distributed architecture.
Wherein, the largest number of subdivision character distributed architectures of character field include 2-4-2 and 1-4-2.First sub- partial character distribution
Structure and the gap of original character distributed architecture are (3-2=1) 1, second sub- partial character distributed architecture and original character point
The gap of cloth structure is (3-1=2) 2.Therefore, first sub- partial character distributed architecture 2-4-2 is determined as reference character distribution
Structure.
Step 2e: the difference character field of above-mentioned original character distributed architecture Yu reference character distributed architecture is determined.
It uses the example above, it is known that the difference character field of original character distributed architecture 3-4-2 and reference character distributed architecture 2-4-2
For first character section (character field on the left of first space).
Step 2f: according to the difference character field, word corresponding with the difference character field in above-mentioned character identification result is corrected
Symbol, obtains revised character identification result.
It uses the example above, since reference character distributed architecture is 2-4-2, when correcting above-mentioned character identification result, can incite somebody to action
Left side first character is left out in character field " 3 " in original character distributed architecture 3-4-2, which is the character more identified.Such as
Fruit reference character distributed architecture is that 3-4-1 can be by original character distributed architecture 3-4- when correcting above-mentioned character identification result
Right side last character is left out in character field " 2 " in 2, which is also the character more identified.
Step 2g: according to above-mentioned space position and revised character identification result, the corresponding character of license plate area is determined
Distributed architecture.
Specifically, step 2g can be executed using the process similar with step 2a.
It is understood that being determined after character identification result is modified according to revised character identification result
The corresponding character distributed architecture of license plate area, can be improved the accuracy of identified character distributed architecture.
As it can be seen that the present embodiment can determine the space position in license plate area according to character identification result, since character is known
Other result can more accurately characterize the information of character in license plate area, therefore the present embodiment can more accurately determine character point
Cloth structure.
In addition, to license plate image carry out character recognition when, difficult point first is that, the shape similar character of kinds of characters type
Symbol can not accurately distinguish.Such as number 1 and letter I, number 0 and letter O, between number 8 and letter b, due to character itself
Shape is very close, easily leads to misrecognition.Also, when the image quality of license plate image is poor, character stroke feature can more mould
Paste, is more difficult to distinguish above-mentioned character in this case.
In order to improve the accuracy of character recognition, after determining the corresponding character distributed architecture of license plate area, this implementation
The method of example can also include the following steps the process of 1~step 3 corrigendum character identification result:
Step 1: judging whether above-mentioned character distributed architecture meets preset characters distributed architecture, if met, execute step
Rapid 2;If do not met, do not handled, i.e., character identification result is not corrected.
Wherein, preset characters distributed architecture are as follows: there are changeless right between character types and the character distributed architecture
The character distributed architecture that should be related to.For example, in European Union area, there are in the every country of 2-3-2 structure, 2-3-2 structure pair
The character types answered are respectively the letter or number of number -2 of 2 letters -3, referring to each with 2-3-2 structure in Fig. 1 b
A license plate image.That is, it is necessarily alphabetical to meet preceding 2 characters in the license plate of 2-3-2 structure, intermediate 3 character certainty
It is number.
Step 2: judging whether the character types in above-mentioned character identification result are preset characters type, if not being default
Character types then follow the steps 3;If it is not handled, it is believed that character identification result is correct, without amendment.
Wherein, preset characters type is character types corresponding with preset characters distributed architecture.
Continuation of the previous cases explanation, it is known that preset characters distributed architecture is 2-3-2, corresponding pre- with the preset characters distributed architecture
If character types are 2 letter or numbers of number -2 of letter -3.If preceding 2 characters are not all in character identification result
Letter, and/or, intermediate 3 characters are not all number, then it is assumed that the character types in above-mentioned character identification result are not predetermined word
Accord with type.
Step 3: the character in above-mentioned character identification result is converted to the word of preset characters type corresponding with the character
Symbol, obtains updated character identification result.
For example, by the character in above-mentioned character identification result according to number 1 and letter I, number 0 and letter O, digital 8 Hes
Corresponding relationship between letter b is converted, it is hereby achieved that updated character identification result.
As it can be seen that the present embodiment can turn the character in character identification result according to determining character distributed architecture
It changes, the different still identical misrecognition characters of shape of character types is corrected, to improve the accurate of character identification result
Property.
In addition, from table 3 it is observed that individual characters distributed architecture corresponds to multiple license plate types.That is, using upper
The target license plate type that the method for stating embodiment determines may more than one.In this case, in order to more accurately determine most
Whole license plate type, can use following implementation.
In a kind of specific embodiment of embodiment shown in Fig. 2, step S205, i.e., according to the target license plate type,
The step of determining the license plate image to be identified corresponding license plate type, can carry out, specifically according to flow diagram shown in Fig. 5
Include the following steps S205a~step S205e:
Step S205a: judging whether the quantity of target license plate type is greater than one, if not, thening follow the steps S205b;
If so, thening follow the steps S205c.
Step S205b: target license plate type is determined as the corresponding license plate type of license plate image to be identified.
Step S205c: according to preset distinctive mark position, distinctive mark region is determined from above-mentioned license plate area.
For example, preset distinctive mark position can be vehicle in the license plate area example of the European Union area shown in Fig. 1 b
The portion in left side in board region, or the region in license plate area between the 2nd character and the 3rd character.Due to
In region in the license plate of European Union area between the 2nd character and the 3rd character, corresponding image has more distinction, therefore is
Raising accuracy, can region between preferred 2nd character and the 3rd character as preset distinctive mark position.
Step S205d: it determines the characteristics of image in the distinctive mark region, and the characteristics of image is inputted into preset be used for really
Determine the classifier of license plate type.The characteristics of image is the feature determined according to the pixel value in above-mentioned distinctive mark region.
Wherein, above-mentioned classifier, characteristics of image obtained and license plate type when for being completed according to the classifier training
Corresponding relationship, determine the corresponding license plate type of above-mentioned inputted characteristics of image.In training, support vector machines can be used
(Support Vector Machine, SVM) trains classifier.
Specifically, histograms of oriented gradients can be used when determining the characteristics of image in the distinctive mark region
(Histogram of Oriented Gradient, HOG) algorithm extracts the characteristics of image in the distinctive mark region.Wherein, HOG
Algorithm be it is a kind of be used to carry out the feature of object detection in image procossing to describe algorithm, by calculating and statistical picture part
The pixel value gradient direction histogram in region carrys out constitutive characteristic.
Step S205e: the license plate type of classifier transmission is obtained, accessed license plate type is determined as to be identified
The corresponding license plate type of license plate image.
As it can be seen that the present embodiment can be when the quantity of target license plate type be greater than one, according to special in license plate area
Mark carries out classification determination using preparatory trained classifier, so as to more accurately determine license plate image to be identified
License plate type.
It should be noted that the recognition methods operand of license plate type provided by the embodiments of the present application is relatively low, it is suitble to
It is applied on the relatively small number of embedded platform of computing resource.
Fig. 6 is a kind of structural schematic diagram of the identification device of license plate type provided by the embodiments of the present application.The device application
In electronic equipment, and it is corresponding with embodiment of the method shown in Fig. 2.The device includes:
Locating module 601 positions the license plate area in the license plate image to be identified for obtaining license plate image to be identified
Domain;
Identification module 602 obtains character identification result for carrying out character recognition to the license plate area;
First determining module 603, for determining the corresponding target word of the license plate area according to the character identification result
Accord with feature;
Matching module 604, for determining the corresponding target carriage of the target character feature from preset license plate typelib
Board type, the license plate typelib, for storing the corresponding relationship of character feature Yu license plate type;
Second determining module 605, for determining that the license plate image to be identified is corresponding according to the target license plate type
License plate type.
In a kind of embodiment of embodiment shown in Fig. 6, the license plate typelib, for storing according to tree-like arrangement
Each layer character feature and the corresponding license plate type of the last layer character feature, each target character feature and the license plate type
Each layer in library corresponds;
The matching module 604 may include:
First determines submodule (not shown), for first layer target character feature to be determined as to the target of current layer
All character features of first layer in license plate typelib are determined as character feature to be selected by character feature;
Matched sub-block (not shown), for the target word with current layer determining from the character feature to be selected
Accord with the matching character feature of characteristic matching;
First judging submodule (not shown), for judging whether the matching character feature is the last layer character
Feature;
Second determines submodule (not shown), for being the last layer character feature when the matching character feature
When, the corresponding license plate type of the matching character feature is determined as target license plate type;
Submodule (not shown) is updated, for when the matching character feature is not the last layer character feature,
Next layer of target character feature is updated to the target character feature of current layer, by next layer all in the license plate typelib
Character feature is updated to character feature to be selected, triggers the matched sub-block.
In a kind of embodiment of embodiment shown in Fig. 6, the target character feature includes character distributed architecture;
First determining module 603 may include:
Space determines submodule (not shown), for determining the license plate area according to the character identification result
In space position;
Structure determination submodule (not shown), for according to each in the space position and the character identification result
A character determines the corresponding character distributed architecture of the license plate area.
In a kind of embodiment of embodiment shown in Fig. 6, the space determines that submodule may include:
First center is away from determination unit (not shown), for determining adjacent character region in the character identification result
Between center away from;
Second center is away from determination unit (not shown), for determining the center away from the maximum preset quantity of middle numerical value
A center away from, as target's center away from;
Mean center away from computing unit, for calculate center of the center away from in addition to the target's center away from other than away from
Average value, as mean center away from;
First judging unit (not shown), for judge the target's center away from the mean center away from whether full
The default size relation of foot;
Space determination unit (not shown), for when the target's center is away from default away from meeting with the mean center
When size relation, the target's center is determined as space position away from the position between corresponding adjacent character region;
Wherein, the default size relation includes at least one of following situations:
The target's center away from the mean center away from ratio be greater than default fractional threshold;
The target's center away from the mean center away from difference be greater than preset difference value threshold value.
In a kind of embodiment of embodiment shown in Fig. 6, described device can also include:
Update module (not shown), for sentencing after determining the corresponding character distributed architecture of the license plate area
Whether the character distributed architecture that breaks meets preset characters distributed architecture;If met, judge in the character identification result
Character types whether be preset characters type, the preset characters type be word corresponding with the preset characters distributed architecture
Accord with type;If being not preset characters type, the character in the character identification result is converted to corresponding with the character
The character of preset characters type obtains updated character identification result.
In a kind of embodiment of embodiment shown in Fig. 6, the structure determination submodule may include:
Originally determined unit (not shown), for according to each in the space position and the character identification result
Character determines original character distributed architecture;
Second judgment unit (not shown) whether there is and the original character in preset structural library for judging
The identical distributed architecture of distributed architecture, wherein the structural library is for storing each character distributed architecture;
First structure determination unit (not shown), for existing in the preset structural library and the original character point
When the identical distributed architecture of cloth structure, the original character distributed architecture is determined as the corresponding character of the license plate area and is distributed
Structure;
Second structure determination unit (not shown), for being not present in the preset structural library and the original character
It is when the identical distributed architecture of distributed architecture, then determining with the original character distributed architecture similarity highest from the structural library
Character distributed architecture, as reference character distributed architecture;Determine that the original character distributed architecture and reference character distribution are tied
The difference character field of structure;According to the difference character field, correct corresponding with the difference character field in the character identification result
Character, obtain revised character identification result;According to the space position and revised character identification result, institute is determined
State the corresponding character distributed architecture of license plate area.
In a kind of embodiment of embodiment shown in Fig. 6, second determining module 605 may include:
Second judgment submodule (not shown), for judging whether the quantity of the target license plate type is greater than one
It is a;
Third determines submodule (not shown), when the quantity of the target license plate type is not more than one, by institute
It states target license plate type and is determined as the corresponding license plate type of the license plate image to be identified;
4th determine submodule (not shown), for when the quantity of the target license plate type be greater than one when, root
According to preset distinctive mark position, distinctive mark region is determined from the license plate area;
Input submodule (not shown), for determining the characteristics of image in the distinctive mark region, and by the figure
As feature inputs the preset classifier for being used to determine license plate type;Wherein, the classifier, for being instructed according to the classifier
Practice the corresponding relationship of characteristics of image and license plate type obtained when completing, the inputted corresponding license plate class of characteristics of image of determination
Type;Described image feature is the feature determined according to the pixel value in the distinctive mark region;
Acquisition submodule (not shown), the license plate type sent for obtaining the classifier, will be accessed
License plate type is determined as the corresponding license plate type of the license plate image to be identified.
Since above-mentioned apparatus embodiment is obtained based on embodiment of the method, and this method technical effect having the same,
Therefore details are not described herein for the technical effect of Installation practice.For device embodiment, since it is substantially similar to method
Embodiment, so describing fairly simple, the relevent part can refer to the partial explaination of embodiments of method.
Fig. 7 is a kind of structural schematic diagram of electronic equipment provided by the embodiments of the present application.The electronic equipment includes processor
701, communication interface 702, memory 703 and communication bus 704.Wherein, processor 701, communication interface 702, memory 703 are logical
It crosses communication bus 704 and completes mutual communication;
Memory 703, for storing computer program;
Processor 701 when for executing the program stored on memory, realizes license plate class provided by the embodiments of the present application
The recognition methods of type.Wherein, the recognition methods of the license plate type includes:
License plate image to be identified is obtained, the license plate area in the license plate image to be identified is positioned;
Character recognition is carried out to the license plate area, obtains character identification result;
According to the character identification result, the corresponding target character feature of the license plate area is determined;
The corresponding target license plate type of the target character feature, the license plate class are determined from preset license plate typelib
Type library, for storing the corresponding relationship of character feature Yu license plate type;
According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.
Wherein, the communication bus that above-mentioned electronic equipment is mentioned can be Peripheral Component Interconnect standard (Peripheral
Component Interconnect, PCI) bus or expanding the industrial standard structure (Extended Industry Standard
Architecture, EISA) bus etc..The communication bus can be divided into address bus, data/address bus, control bus etc..For just
It is only indicated with a thick line in expression, figure, it is not intended that an only bus or a type of bus.
Communication interface is for the communication between above-mentioned electronic equipment and other equipment.
Memory may include random access memory (Random Access Memory, RAM), also may include non-easy
The property lost memory (Non-Volatile Memory, NVM), for example, at least a magnetic disk storage.Optionally, memory may be used also
To be storage device that at least one is located remotely from aforementioned processor.
Above-mentioned processor can be general processor, including central processing unit (Central Processing Unit,
CPU), network processing unit (Network Processor, NP) etc.;It can also be digital signal processor (Digital Signal
Processing, DSP), it is specific integrated circuit (Application Specific Integrated Circuit, ASIC), existing
It is field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic device, discrete
Door or transistor logic, discrete hardware components.
As it can be seen that the present embodiment can determine license plate type according to the character feature of license plate.In the vehicle of different license plate types
In the identical situation of board colouring information, the character feature of license plate is the feature more richer than the colouring information of license plate, according to license plate
Character feature can more accurately identify license plate type.Meanwhile character recognition is carried out to license plate area, and according to character recognition
As a result determine license plate character feature, determining character feature can be made more acurrate, thus further increase the present embodiment provides
License plate type identification scheme accuracy rate.
The embodiment of the present application also provides a kind of computer readable storage medium, stored in the computer readable storage medium
There is computer program, the identification side of license plate type provided by the embodiments of the present application is realized when computer program is executed by processor
Method.Wherein, the recognition methods of the license plate type includes:
License plate image to be identified is obtained, the license plate area in the license plate image to be identified is positioned;
Character recognition is carried out to the license plate area, obtains character identification result;
According to the character identification result, the corresponding target character feature of the license plate area is determined;
The corresponding target license plate type of the target character feature, the license plate class are determined from preset license plate typelib
Type library, for storing the corresponding relationship of character feature Yu license plate type;
According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.
As it can be seen that the present embodiment can determine license plate type according to the character feature of license plate.In the vehicle of different license plate types
In the identical situation of board colouring information, the character feature of license plate is the feature more richer than the colouring information of license plate, according to license plate
Character feature can more accurately identify license plate type.Meanwhile character recognition is carried out to license plate area, and according to character recognition
As a result determine license plate character feature, determining character feature can be made more acurrate, thus further increase the present embodiment provides
License plate type identification scheme accuracy rate.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or any other variant be intended to it is non-
It is exclusive to include, so that the process, method, article or equipment for including a series of elements not only includes those elements,
It but also including other elements that are not explicitly listed, or further include solid by this process, method, article or equipment
Some elements.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including
There is also other identical elements in the process, method, article or equipment of the element.
Each embodiment in this specification is all made of relevant mode and describes, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.
The foregoing is merely the preferred embodiments of the application, are not intended to limit the protection scope of the application.It is all
Any modification, equivalent substitution, improvement and etc. done within spirit herein and principle are all contained in the protection scope of the application
It is interior.
Claims (16)
1. a kind of recognition methods of license plate type, which is characterized in that the described method includes:
License plate image to be identified is obtained, the license plate area in the license plate image to be identified is positioned;
Character recognition is carried out to the license plate area, obtains character identification result;
According to the character identification result, the corresponding target character feature of the license plate area is determined;
The corresponding target license plate type of the target character feature, the license plate type are determined from preset license plate typelib
Library, for storing the corresponding relationship of character feature Yu license plate type;
According to the target license plate type, the corresponding license plate type of the license plate image to be identified is determined.
2. the method according to claim 1, wherein the license plate typelib, is used to store according to tree-like arrangement
Each layer character feature and the corresponding license plate type of the last layer character feature, each target character feature and the license plate class
Each layer in type library corresponds;
The described the step of corresponding target license plate type of the target character feature is determined from preset license plate typelib, packet
It includes:
First layer target character feature is determined as to the target character feature of current layer, by first layer in the license plate typelib
All character features are determined as character feature to be selected;
The determining matching character feature with the target character characteristic matching of current layer from the character feature to be selected;
Judge whether the matching character feature is the last layer character feature;
If it is, the corresponding license plate type of the matching character feature is determined as target license plate type;
If it is not, then next layer of target character feature to be updated to the target character feature of current layer, by the license plate typelib
In next layer of all character features be updated to character feature to be selected, return execute it is described from the character feature to be selected
The step of determining matching character feature with the target character characteristic matching of current layer.
3. described in any item methods according to claim 1~2, which is characterized in that the target character feature includes following spy
At least one of sign: single bilayer license board information, character total quantity, the quantity of all kinds of ocra font ocrs, character distributed architecture.
4. according to the method described in claim 3, it is characterized in that, the target character feature includes the character distribution knot
Structure;
It is described according to the character identification result, the step of determining the license plate area corresponding target character feature, comprising:
According to the character identification result, the space position in the license plate area is determined;
According to each character in the space position and the character identification result, the corresponding character point of the license plate area is determined
Cloth structure.
5. according to the method described in claim 4, determining the vehicle it is characterized in that, described according to the character identification result
The step of space position in board region, comprising:
Determine center in the character identification result between adjacent character region away from;
Determine the center away from the maximum preset quantity center of middle numerical value away from, as target's center away from;
Calculate center of the center away from in addition to the target's center away from other than away from average value, as mean center away from;
Judge the target's center away from the mean center away from whether meeting default size relation;
If it is satisfied, then the target's center is determined as space position away from the position between corresponding adjacent character region;
Wherein, the default size relation includes at least one of following situations:
The target's center away from the mean center away from ratio be greater than default fractional threshold;
The target's center away from the mean center away from difference be greater than preset difference value threshold value.
6. according to the method described in claim 4, it is characterized in that, determining the corresponding character distributed architecture of the license plate area
Later, the method also includes:
Judge whether the character distributed architecture meets preset characters distributed architecture;
If met, judge whether the character types in the character identification result are preset characters type, the predetermined word
Symbol type is character types corresponding with the preset characters distributed architecture;
If not being preset characters type, the character in the character identification result is converted to corresponding with the character default
The character of character types obtains updated character identification result.
7. according to the method described in claim 4, it is characterized in that, described according to the space position and the character recognition knot
Each character in fruit, the step of determining the license plate area corresponding character distributed architecture, comprising:
According to each character in the space position and the character identification result, original character distributed architecture is determined;
Judge in preset structural library with the presence or absence of distributed architecture identical with the original character distributed architecture, wherein described
Structural library is for storing each character distributed architecture;
If it is present the original character distributed architecture is determined as the corresponding character distributed architecture of the license plate area;
If it does not exist, then determining from the structural library be distributed with the highest character of the original character distributed architecture similarity
Structure, as reference character distributed architecture;
Determine the difference character field of the original character distributed architecture Yu reference character distributed architecture;
According to the difference character field, character corresponding with the difference character field in the character identification result is corrected, is obtained
Revised character identification result;
According to the space position and revised character identification result, the corresponding character distribution knot of the license plate area is determined
Structure.
8. the method according to claim 1, wherein described according to the target license plate type, determine it is described to
The step of identifying license plate image corresponding license plate type, comprising:
Judge whether the quantity of the target license plate type is greater than one;
If it is not, then the target license plate type is determined as the corresponding license plate type of the license plate image to be identified;
If it is, determining distinctive mark region from the license plate area according to preset distinctive mark position;
Determine the characteristics of image in the distinctive mark region, and be used to determine license plate type for the input of described image feature is preset
Classifier;Wherein, the classifier, characteristics of image obtained and license plate class when for being completed according to the classifier training
The corresponding relationship of type determines the inputted corresponding license plate type of characteristics of image;Described image feature is according to the special mark
The feature that the pixel value in will region determines;
The license plate type that the classifier is sent is obtained, accessed license plate type is determined as the license plate image to be identified
Corresponding license plate type.
9. a kind of identification device of license plate type, which is characterized in that described device includes:
Locating module positions the license plate area in the license plate image to be identified for obtaining license plate image to be identified;
Identification module obtains character identification result for carrying out character recognition to the license plate area;
First determining module, for determining the corresponding target character feature of the license plate area according to the character identification result;
Matching module, for determining the corresponding target license plate type of the target character feature from preset license plate typelib,
The license plate typelib, for storing the corresponding relationship of character feature Yu license plate type;
Second determining module, for determining the corresponding license plate class of the license plate image to be identified according to the target license plate type
Type.
10. device according to claim 9, which is characterized in that the license plate typelib, for storing according to tree-like arrangement
Each layer character feature and the corresponding license plate type of the last layer character feature, each target character feature and the license plate class
Each layer in type library corresponds;
The matching module, comprising:
First determines submodule, will be described for first layer target character feature to be determined as to the target character feature of current layer
All character features of first layer are determined as character feature to be selected in license plate typelib;
Matched sub-block, for the matching with the target character characteristic matching of current layer determining from the character feature to be selected
Character feature;
First judging submodule, for judging whether the matching character feature is the last layer character feature;
Second determines submodule, is used for when the matching character feature is the last layer character feature, by the matching character
The corresponding license plate type of feature is determined as target license plate type;
Submodule is updated, for when the matching character feature is not the last layer character feature, by next layer of target character
Feature is updated to the target character feature of current layer, by next layer in the license plate typelib of all character features be updated to
The character feature of choosing triggers the matched sub-block.
11. according to the described in any item devices of claim 9~10, which is characterized in that the target character feature includes character
Distributed architecture;
First determining module, comprising:
Space determines submodule, for determining the space position in the license plate area according to the character identification result;
Structure determination submodule, described in determining according to each character in the space position and the character identification result
The corresponding character distributed architecture of license plate area.
12. device according to claim 11, which is characterized in that the space determines submodule, comprising:
First center away from determination unit, for determine the center in the character identification result between adjacent character region away from;
Second center is away from determination unit, for determining the center away from the maximum preset quantity center of middle numerical value away from as mesh
Mark center away from;
Mean center away from computing unit, for calculate center of the center away from in addition to the target's center away from other than away from it is flat
Mean value, as mean center away from;
First judging unit, for judge the target's center away from the mean center away from whether meeting default size relation;
Space determination unit is used for when the target's center is away from the mean center away from default size relation is met, by institute
It states target's center and is determined as space position away from the position between corresponding adjacent character region;
Wherein, the default size relation includes at least one of following situations:
The target's center away from the mean center away from ratio be greater than default fractional threshold;
The target's center away from the mean center away from difference be greater than preset difference value threshold value.
13. device according to claim 11, which is characterized in that described device further include:
Update module, for after determining the corresponding character distributed architecture of the license plate area, judging the character distribution knot
Whether structure meets preset characters distributed architecture;If met, judge the character types in the character identification result whether be
Preset characters type, the preset characters type are character types corresponding with the preset characters distributed architecture;If not being
Character in the character identification result is then converted to the word of preset characters type corresponding with the character by preset characters type
Symbol, obtains updated character identification result.
14. device according to claim 11, which is characterized in that the structure determination submodule, comprising:
Originally determined unit, for determining initial word according to each character in the space position and the character identification result
Accord with distributed architecture;
Second judgment unit whether there is and identical point of the original character distributed architecture in preset structural library for judging
Cloth structure, wherein the structural library is for storing each character distributed architecture;
There is distribution identical with the original character distributed architecture in preset structural library for working as in first structure determination unit
When structure, the original character distributed architecture is determined as the corresponding character distributed architecture of the license plate area;
Second structure determination unit, for being not present in the preset structural library and identical point of the original character distributed architecture
It is when cloth structure, then determining with the highest character distributed architecture of the original character distributed architecture similarity from the structural library,
As reference character distributed architecture;Determine the difference character field of the original character distributed architecture Yu reference character distributed architecture;
According to the difference character field, character corresponding with the difference character field in the character identification result is corrected, is corrected
Character identification result afterwards;According to the space position and revised character identification result, determine that the license plate area is corresponding
Character distributed architecture.
15. device according to claim 9, which is characterized in that second determining module, comprising:
Second judgment submodule, for judging whether the quantity of the target license plate type is greater than one;
Third determines submodule, when the quantity of the target license plate type is not more than one, the target license plate type is true
It is set to the corresponding license plate type of the license plate image to be identified;
4th determine submodule, for when the quantity of the target license plate type be greater than one when, according to preset distinctive mark
Position determines distinctive mark region from the license plate area;
Input submodule, for determining the characteristics of image in the distinctive mark region, and the input of described image feature is preset
For determining the classifier of license plate type;Wherein, the classifier, it is obtained when for being completed according to the classifier training
The corresponding relationship of characteristics of image and license plate type determines the inputted corresponding license plate type of characteristics of image;Described image feature
For the feature determined according to the pixel value in the distinctive mark region;
Acquisition submodule, the license plate type sent for obtaining the classifier, is determined as institute for accessed license plate type
State the corresponding license plate type of license plate image to be identified.
16. a kind of electronic equipment, which is characterized in that including processor, communication interface, memory and communication bus, wherein processing
Device, communication interface, memory complete mutual communication by communication bus;
Memory, for storing computer program;
Processor when for executing the program stored on memory, realizes any method and step of claim 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710946300.2A CN109657664B (en) | 2017-10-12 | 2017-10-12 | License plate type identification method and device and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710946300.2A CN109657664B (en) | 2017-10-12 | 2017-10-12 | License plate type identification method and device and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109657664A true CN109657664A (en) | 2019-04-19 |
CN109657664B CN109657664B (en) | 2021-02-05 |
Family
ID=66109282
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710946300.2A Active CN109657664B (en) | 2017-10-12 | 2017-10-12 | License plate type identification method and device and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109657664B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110728276A (en) * | 2018-07-16 | 2020-01-24 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device |
CN111652222A (en) * | 2020-07-13 | 2020-09-11 | 深圳市智搜信息技术有限公司 | License plate positioning method and device, computer equipment and storage medium |
CN111860539A (en) * | 2020-07-20 | 2020-10-30 | 济南博观智能科技有限公司 | License plate color recognition method, device and medium |
CN112598005A (en) * | 2020-12-24 | 2021-04-02 | 深圳市信义科技有限公司 | Character segmentation method for license plate of armed police, storage medium and computer equipment |
CN112712626A (en) * | 2019-10-24 | 2021-04-27 | 上海际链网络科技有限公司 | Vehicle verification method and device based on license plate information, computer equipment and storage medium |
CN112861567A (en) * | 2019-11-12 | 2021-05-28 | 上海高德威智能交通系统有限公司 | Vehicle type classification method and device |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102509458A (en) * | 2011-10-19 | 2012-06-20 | 杭州华三通信技术有限公司 | Fake-licensed vehicle identifying method and device |
CN102693431A (en) * | 2012-05-31 | 2012-09-26 | 信帧电子技术(北京)有限公司 | Method and device for identifying type of white number plate |
CN102722733A (en) * | 2012-05-31 | 2012-10-10 | 信帧电子技术(北京)有限公司 | Identification method and device of license plate types |
US20130182910A1 (en) * | 2012-01-18 | 2013-07-18 | Xerox Corporation | License plate optical character recognition method and system |
CN104123302A (en) * | 2013-04-27 | 2014-10-29 | 四川火狐无线科技有限公司 | Searching method, device and system |
US20150324654A1 (en) * | 2013-04-11 | 2015-11-12 | International Business Machines Corporation | Determining images having unidentifiable license plates |
CN105224956A (en) * | 2015-10-16 | 2016-01-06 | 深圳市捷顺科技实业股份有限公司 | A kind of car plate classification based on multi-feature fusion recognition methods and system |
CN105589862A (en) * | 2014-10-21 | 2016-05-18 | 杭州华为企业通信技术有限公司 | License plate data index structure building method, retrieval method and device |
CN105631470A (en) * | 2015-12-21 | 2016-06-01 | 深圳市捷顺科技实业股份有限公司 | Method and system for verifying license plate type |
-
2017
- 2017-10-12 CN CN201710946300.2A patent/CN109657664B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102509458A (en) * | 2011-10-19 | 2012-06-20 | 杭州华三通信技术有限公司 | Fake-licensed vehicle identifying method and device |
US20130182910A1 (en) * | 2012-01-18 | 2013-07-18 | Xerox Corporation | License plate optical character recognition method and system |
CN102693431A (en) * | 2012-05-31 | 2012-09-26 | 信帧电子技术(北京)有限公司 | Method and device for identifying type of white number plate |
CN102722733A (en) * | 2012-05-31 | 2012-10-10 | 信帧电子技术(北京)有限公司 | Identification method and device of license plate types |
US20150324654A1 (en) * | 2013-04-11 | 2015-11-12 | International Business Machines Corporation | Determining images having unidentifiable license plates |
CN104123302A (en) * | 2013-04-27 | 2014-10-29 | 四川火狐无线科技有限公司 | Searching method, device and system |
CN105589862A (en) * | 2014-10-21 | 2016-05-18 | 杭州华为企业通信技术有限公司 | License plate data index structure building method, retrieval method and device |
CN105224956A (en) * | 2015-10-16 | 2016-01-06 | 深圳市捷顺科技实业股份有限公司 | A kind of car plate classification based on multi-feature fusion recognition methods and system |
CN105631470A (en) * | 2015-12-21 | 2016-06-01 | 深圳市捷顺科技实业股份有限公司 | Method and system for verifying license plate type |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110728276A (en) * | 2018-07-16 | 2020-01-24 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device |
CN112712626A (en) * | 2019-10-24 | 2021-04-27 | 上海际链网络科技有限公司 | Vehicle verification method and device based on license plate information, computer equipment and storage medium |
CN112861567A (en) * | 2019-11-12 | 2021-05-28 | 上海高德威智能交通系统有限公司 | Vehicle type classification method and device |
CN112861567B (en) * | 2019-11-12 | 2024-01-05 | 上海高德威智能交通系统有限公司 | Vehicle type classification method and device |
CN111652222A (en) * | 2020-07-13 | 2020-09-11 | 深圳市智搜信息技术有限公司 | License plate positioning method and device, computer equipment and storage medium |
CN111860539A (en) * | 2020-07-20 | 2020-10-30 | 济南博观智能科技有限公司 | License plate color recognition method, device and medium |
CN112598005A (en) * | 2020-12-24 | 2021-04-02 | 深圳市信义科技有限公司 | Character segmentation method for license plate of armed police, storage medium and computer equipment |
Also Published As
Publication number | Publication date |
---|---|
CN109657664B (en) | 2021-02-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109657664A (en) | A kind of recognition methods, device and the electronic equipment of license plate type | |
US11455805B2 (en) | Method and apparatus for detecting parking space usage condition, electronic device, and storage medium | |
CN107194398B (en) | Vehicle damages recognition methods and the system at position | |
CN106599792B (en) | Method for detecting hand driving violation behavior | |
CN105574550A (en) | Vehicle identification method and device | |
CN109740424A (en) | Traffic violations recognition methods and Related product | |
CN105975929A (en) | Fast pedestrian detection method based on aggregated channel features | |
CN110111581A (en) | Target identification method, device, computer equipment and storage medium | |
CN105631470A (en) | Method and system for verifying license plate type | |
CN104766042A (en) | Method and apparatus for and recognizing traffic sign board | |
CN105512660A (en) | License number identification method and device | |
CN108090429A (en) | Face bayonet model recognizing method before a kind of classification | |
CN103226696B (en) | The identification system and method for car plate | |
CN103164701B (en) | Handwritten Numeral Recognition Method and device | |
CN105354533B (en) | A kind of unlicensed vehicle model recognizing method of bayonet based on bag of words | |
CN103544480A (en) | Vehicle color recognition method | |
CN105631445A (en) | Character recognition method and system for license plate with Chinese characters | |
CN108764302A (en) | A kind of bill images sorting technique based on color characteristic and bag of words feature | |
CN110909598A (en) | Deep learning-based method for recognizing illegal traffic driving of non-motor vehicle lane | |
CN105184291A (en) | Method and system for detecting multiple types of license plates | |
CN105160292A (en) | Vehicle identification recognition method and system | |
CN111860219B (en) | High-speed channel occupation judging method and device and electronic equipment | |
CN104794432A (en) | Method and system for rapid cascade type car logo vision inspection and identification | |
CN111008554B (en) | Deep learning-based method for identifying pedestrians without giving away in dynamic traffic zebra stripes | |
CN104156701A (en) | Plate number similar character recognition method based on decision-making tree and SVM |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |